OpenAI launches GPT-6 Sol and Luna as cheaper models after Astra

OpenAI launched GPT-6 Sol and GPT-6 Luna today, new models following GPT-6 Astra. The company says they improve professional work, factuality, coding, computer use and alignment.

GPT-6 Sol is designed for everyday tasks and has greater reasoning capability. GPT-6 Luna is cheaper and optimized for fast responses and higher-volume use.

Benchmark results and response style

OpenAI says both models outperform their predecessors on its internal tests. The company says Sol makes about half as many mistakes as GPT-5.6 Sol and also improves on coding and computer-use evaluations.

GPT-6 Sol matches or exceeds Claude Fable 5.1 on some benchmarks. OpenAI also compared Sol with Opus 5, but Anthropic released Opus 5.5 on the same day and says it outperforms Astra on some coding and knowledge-work tests.

The models adopt Astra’s communication style, with shorter answers, less jargon and fewer low-value details. OpenAI also found lower rates of misleading claims about coding work in its alignment testing.

Token prices and prompt caching

GPT-6 Sol costs $2 per million input tokens and $10 per million output tokens, half the price of GPT-5.6 Sol. GPT-6 Luna costs $0.10 and $0.50 respectively, down from $0.20 and $1.20 for its predecessor.

OpenAI says improved prompt caching will let agents reuse more context, respond faster and receive a 90% discount on cached input-token reads. The company notes that Sol’s price is not a direct comparison with Opus 5.5 because Anthropic says its model uses fewer tokens per task than Opus 5.

Availability of the new models

GPT-6 Sol and GPT-6 Luna are rolling out today in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users. Luna is also available to Free and Go users in the desktop app.

Neither model is currently available in Chat mode.